There are a number of ways you can approach this. Some of it depends on what you need to comply to.
The first thing is that if those calibrated thermocouples must be tracked, you need to identify when ever you replace an old with a new. Therefore, you probably could not get away with the same ID number on new ones. However, as an alternative, if you document in your history when ever an old is replaced by a new, and the surrounding details (calibration data, etc.) are captured, it may be acceptable to use the same ID number. And for audit purposes, having captured everytime you replace a thermocouple may suffice.
Some of the above depends on how long your normal calibration interval is, and how often you have to replace broken thermocouples.
We actually have to calibrate type R thermocouple sets, and put new wire in them every calibration (as the wire degrades in the very high temp environment).
As for the recall dates, no matter what your calibration interval is, you can always recall items early. For example if you had to replace a thermocouple six months into a twelve month interval, and all the rest of the thermocouples are due in six months, you can reasonably and legitimately recall the replaced one early to put it back in line with the due dates on the rest of those on the same machine.
As for your other question about issuing a thermocouple as a spare, that is actually what we do. We have sets of thermocouples in a sheath with an ID number for that set. We calibrate them, then start the calibration interval when we issue them (as it is presumed through our experience that the thermocouple does not deteriorate until we put it into the machine). The important part of how ever you decide to set up your ID numbering is that you be able to somehow positively connect a specific thermocouple with specific calibration information. I do suggest that if you use different thermocouple wire with the same ID, you somehow track every time you change wire.
